Keithley 2000 Calibration Window

[timbocgn]

Hi guys,

I had the chance to test my repaired K2000 (last cal 1998) against a Fluke Calibrator, but there was no time to do a real calibration cycle.

I have now the measurement readings for most of the ranges and I am wondering if I have to check some ranges for hardware defects (I found some during my repair) again since they are "off".

Does anyone know the calibration window of the ranges, so how much deviation can be compensated / is tolarated by the calibration?

Next time I have a chance I would not like to end up with the result that a calibration is not possible....I would like to fix these potential flaw upfront.

Cheers,

Tim.

[voltsandjolts]

Does anyone know the calibration window of the ranges, so how much deviation can be compensated / is tolarated by the calibration?

The calibration manual explains process and limits of applied stimulus.
https://www.tek.com/en/manual/digital-multimeter/broad-purpose-digital-multimeters-manual-6-keithley-2000-series-6-digit-multimeter-scanning

[timbocgn]

Hi voltsandjolts,

I read the manual again, but I am not sure if I just missed the information.

The calibration manual gives me all the tolearance for the ranges and also offers ranges of nominal voltages (in case the calibrator cannot produce the default voltages).

However I did not find values for the ranges which the calibration algorithm accepts as „compensateable“ so to say:

Imagine I calibrate the 10V range and I apply 8V (or my K2000 is defect and thinks it is 8V), would the compensation then adjust the DAC measurements to match the difference?

For small values this is exactly the idea of the calibration, but what is the maximum difference the algorithm tolerates?

Thanks in advance for helping out!

Tim.

[benj38]

To answer your question (I hope): I have never seen any instrument for which the information you are looking for is provided, and for a good reason.

Obviously, at the basis of the calibration process is the assumption that the instrument has no faults, and is simply misaligned. Thus, all the manufacturer is "specifying" about the calibration "ranges of compensation" is that they are enough to get the instrument back into spec if it has no faults. In particular, the information in table 2.8 in the cal manual of K2000 is not about the "ranges of compensation" but about the flexibility one has in choosing the calibration equipment and its settings.

The reason the manufacturers do not give the "ranges of compensation" is that this information is considered useless: if the instrument has no faults then all one needs to know is that these ranges are enough to fully calibrate it; and if the instrument has a fault then, no matter what these ranges are, one cannot assume that by adjusting things during calibration the instrument will meet specs.
For example, assume that your K2000 has a fault which negatively affects the linearity of the DAC at part of its range. In that case, even if you can adjust things during calibration such that the meter correctly reads the voltages used during calibration, some intermediate values between these will be out of spec., and no amount of "calibration" would fix that.

It is the purpose of performance tests -- which usually exercise the instrument over many more points of operation than the calibration process does -- to allow one to fairly confidently assume that an instrument is working properly (but even that is not guaranteed if some more insidious fault exists).